At last, serving huge quantities of 3D geospatial data into interactive user applications is getting easier and more accessible. Support for ‘3D streaming’ is gaining a foothold within popular geospatial applications, paving the way for data providers to present their own 3D data assets in ways that users will find natural and accessible. For us at OS, this is a development that we’re very excited about, as we believe it will be a powerful tool in helping to unlock the potential of geospatial 3D data.
For many people, Google Earth (launched in 2005) was their first experience of navigating a ‘digital globe’ – a tool that is typically supported by ‘streaming’ technology. Overwhelming volumes of geospatial information were clearly available, yet how had it got there? Although Google Earth was then only available as a downloadable app, it was clear that the data was being sourced quite independently. For many of us, this served as our introduction to ‘3D streaming’ – the ability to selectively deliver content, based upon real-time navigation, within a 3D scene.
Since that time, things have developed apace. Google Earth has developed further, with new types of content, including large expanses of 3D cityscapes. Alongside that, Google Maps, is also now able to stream similar 3D content. Other suppliers now offer equivalent 3D services, with Apple Maps perhaps being the best known of these. Until recently, though, the facility to publish data using 3D streaming functionality has not been easily accessible to the wider GI user and the developer community. This situation is now being rectified, with the emergence of two new open 3D streaming formats – Indexed 3D Scene Layers and 3D Tiles.
Figure 1- A point cloud of Bournemouth, streamed in i3s format via ESRI’s ArcGIS Online
‘Indexed 3D Scene Layers’, developed by ESRI and commonly referred to as just ‘i3s’, was first released in 2015. Since then, support for it has been built into many of ESRI’s own products. Meanwhile, ‘3D Tiles’ (sometimes also known as ‘Cesium 3D Tiles’) is a broadly similar standard, initially created by AGI for use within the Cesium virtual globe – a technology used within our popular OS Maps service. Currently, both 3D Tiles and i3s are candidate OGC community standards and have attracted support from third party application developers.
How OS is using i3s
Of these two formats, we have chosen, initially, to concentrate on i3s, simply due to the levels of application support currently in place. i3s supports several common types of 3D geospatial data, including ‘3D objects’, ‘integrated meshes’ and point clouds – with scope for more to be added in later versions. In many cases, you can also add individual identities and ‘key/value pair’ attribution to the 3D assets. These data are then segmented and indexed – at different levels of detail – before typically being persisted as an i3s ‘package’ file. The files can then be read by a supporting 3D client application (e.g. ArcGIS Online), where they may be used to support dynamic and compelling user experiences. Despite the apparent complexity in subject matter, producing and using i3s data has proven to be very straightforward, with much of the supporting functionality kept hidden from users.
Figure 2 – Extruded OS MasterMap building footprints, PointX and AddressBase, all streamed into an ArcGIS Online app.
To stimulate interest, we’ve already built several i3s demonstrators that present our own data products within interactive scenes. In one example, shown above, we used an extensive coverage of OS MasterMap Topography Layer area building footprints, heighted using OS Terrain 50, then converted into 3D objects using our Building Height Attribute dataset. Alongside this, we included symbolised representations of our Points of Interest and AddressBase products, the 3D symbols themselves being created separately within Blender. With so much 3D content present, such a scene would have previously been prone to performance limitations. However, by making use of ESRI’s i3s support, our client application (in this case, created via ArcGIS Online) could deliver a successful and immersive 3D experience.
Within OS, we’re certainly excited about the potential of this technology and look forward to discovering how our customers will incorporate OS data within their own streamed 3D worlds. So, has your organisation started to experiment with 3D streaming? Is this something that you feel would be of interest to you in the future? We’d love to hear your thoughts…